Re: [PATCH] FM1216ME_MK3 some changes
Am Donnerstag, den 14.05.2009, 03:10 +0200 schrieb hermann pitton: Hi Dmitry :) [snip] That depends on how up to date the app is, once selected, even per channel, you don't have to do this ever again and don't need any insmod options at all currently. I recently tried to keep it for others the same :) Umm, for the analog only FM1216ME/ I H-3 (MK3) it seems you have some other stuff on it employed. On a first look ;) Dmitry, any news on that? I don't want to hinder you to do improvements, but what about this SAW filters? Are they really approved to have a Philips label on the tuner? The hand/foot ;) painted claimed performance statistics don't look like that. Am I wrong? Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi 1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. for 2.: Discussed to the end if you stay at 441MHz. If you still want to have it in, just send a patch and no more info is needed. (Likely Andy is giving only examples for more difficult cases, sorry.) for 1.: I would like to be absolutely sure, that we are talking about the same tuner. I want to have the exact filters on it at least. I would also say that, if we need to implement AGC TOP control, it would be better to add it at struct v4l2_tuner (VIDIOC_G_TUNER and VIDIOC_S_TUNER), instead of adding it as a control. Good idea. For MK3 need control AGC TOP and TOP of TDA9887 For MK5 need control only TOP of TDA9887 With my best regards, Dmitry. 3.: That is what Andy noted. Following the Philips datasheet for TOP, it should be added for negative modulation, positive modulation and FM accordingly. (2 and 3 are out of discussion) If you still have some sort of Secam fire and can improve it, we must know the tuner you are on exactly. If it is the original Philips, without any such TOP suggestions over the full range in recent datasheets (???), I assume you might have them, I would say you can proceed, if you have shown that you are really still on the same tuner. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi Dmitry, Am Mittwoch, den 13.05.2009, 07:37 +1000 schrieb Dmitri Belimov: Hi hermann Am Sonntag, den 10.05.2009, 08:52 +1000 schrieb Dmitri Belimov: Hi All. [snip] Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. NO! it is not Dmitri, can you cut one off and tell us what it is all about ? Unless you do so, all other is pointless and I likely stop to participate in such stuff. Sorry my delay. I lost subject of discussion. What main question?? 1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. for 2.: Discussed to the end if you stay at 441MHz. If you still want to have it in, just send a patch and no more info is needed. (Likely Andy is giving only examples for more difficult cases, sorry.) for 1.: I would like to be absolutely sure, that we are talking about the same tuner. I want to have the exact filters on it at least. 3.: That is what Andy noted. Following the Philips datasheet for TOP, it should be added for negative modulation, positive modulation and FM accordingly. (2 and 3 are out of discussion) If you still have some sort of Secam fire and can improve it, we must know the tuner you are on exactly. If it is the original Philips, without any such TOP suggestions over the full range in recent datasheets (???), I assume you might have them, I would say you can proceed, if you have shown that you are really still on the same tuner. This is real Philips MK3 and MK5 tuner. We have all docs from vendor. This is some photos of the MK3 hybrid tuner. I make photos of the analog MK3 little later. thanks a lot. I had only a first look on this one. There are a lot of different revisions, yours is F4 on the tuner PCB. Taifun PLL versions vary, different combinations of ceramic and SMD filters are used for the radio IF, but even on a revision C1 with Taifun 6034T A1 and tuner revision SV21 0437 (CTX918_V2 md7134) all three SAW filters are the same already. A CTX925 with your Taifun 6034T B1 had also no better performance for what I can say. PCB still has C1, but SV21 0502 tuner revision. Important to know was, that you are on original Philips tuners and SAW filters are the same. Guess this will hold for the analog FM1216ME/I MK3 too. Good decision, it are still the best tuners for analog and digital I have around, but can't talk for Secam D/K. I wonder about missing complaints during the last four years. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. for 2.: Discussed to the end if you stay at 441MHz. If you still want to have it in, just send a patch and no more info is needed. (Likely Andy is giving only examples for more difficult cases, sorry.) for 1.: I would like to be absolutely sure, that we are talking about the same tuner. I want to have the exact filters on it at least. I would also say that, if we need to implement AGC TOP control, it would be better to add it at struct v4l2_tuner (VIDIOC_G_TUNER and VIDIOC_S_TUNER), instead of adding it as a control. 3.: That is what Andy noted. Following the Philips datasheet for TOP, it should be added for negative modulation, positive modulation and FM accordingly. (2 and 3 are out of discussion) If you still have some sort of Secam fire and can improve it, we must know the tuner you are on exactly. If it is the original Philips, without any such TOP suggestions over the full range in recent datasheets (???), I assume you might have them, I would say you can proceed, if you have shown that you are really still on the same tuner. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
On Mon, 11 May 2009 05:35:51 +0200 hermann pitton hermann-pit...@arcor.de wrote: Dmitri, Am Sonntag, den 10.05.2009, 08:52 +1000 schrieb Dmitri Belimov: Hi All. [snip] Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. NO! it is not Dmitri, can you cut one off and tell us what it is all about ? Unless you do so, all other is pointless and I likely stop to participate in such stuff. Sorry my delay. I lost subject of discussion. What main question?? 1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. for 2.: Discussed to the end if you stay at 441MHz. If you still want to have it in, just send a patch and no more info is needed. (Likely Andy is giving only examples for more difficult cases, sorry.) This is my patch. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com for 1.: I would like to be absolutely sure, that we are talking about the same tuner. I want to have the exact filters on it at least. 3.: That is what Andy noted. Following the Philips datasheet for TOP, it should be added for negative modulation, positive modulation and FM accordingly. (2 and 3 are out of discussion) If you still have some sort of Secam fire and can improve it, we must know the tuner you are on exactly. If it is the original Philips, without any such TOP suggestions over the full range in recent datasheets (???), I assume you might have them, I would say you can proceed, if you have shown that you are really still on the same tuner. Ok. I read the datasheet, make photos inside tuner. With my best regards, Dmitry. Cheers, Hermann diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com
Re: [PATCH] FM1216ME_MK3 some changes
Hi All. [snip] Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. NO! it is not Dmitri, can you cut one off and tell us what it is all about ? Unless you do so, all other is pointless and I likely stop to participate in such stuff. Sorry my delay. I lost subject of discussion. What main question?? 1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. With my best regards, Dmitry. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Dmitri, Am Sonntag, den 10.05.2009, 08:52 +1000 schrieb Dmitri Belimov: Hi All. [snip] Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. NO! it is not Dmitri, can you cut one off and tell us what it is all about ? Unless you do so, all other is pointless and I likely stop to participate in such stuff. Sorry my delay. I lost subject of discussion. What main question?? 1. AGC TOP of RF part - I think need support for MK3 2. Changing to 441MHz is not critical. We can write some information about this case to Wiki or docs. for 2.: Discussed to the end if you stay at 441MHz. If you still want to have it in, just send a patch and no more info is needed. (Likely Andy is giving only examples for more difficult cases, sorry.) for 1.: I would like to be absolutely sure, that we are talking about the same tuner. I want to have the exact filters on it at least. 3.: That is what Andy noted. Following the Philips datasheet for TOP, it should be added for negative modulation, positive modulation and FM accordingly. (2 and 3 are out of discussion) If you still have some sort of Secam fire and can improve it, we must know the tuner you are on exactly. If it is the original Philips, without any such TOP suggestions over the full range in recent datasheets (???), I assume you might have them, I would say you can proceed, if you have shown that you are really still on the same tuner. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
[snip] Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. NO! it is not Dmitri, can you cut one off and tell us what it is all about ? Unless you do so, all other is pointless and I likely stop to participate in such stuff. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
On Fri, 2009-05-08 at 02:08 +0200, hermann pitton wrote: Am Mittwoch, den 06.05.2009, 23:03 -0400 schrieb Andy Walls: For the change of UHF start I don't see any problem. If you're talking about the frequency for the bandswitch, I don't see a problem either in general. It may cause a problem for clones of the FM1216ME MK3 that don't have the same filter performance near the cutover, but use the same tuner defintion as the FM1216ME MK3 in tuner-types.c. It may be best to point any clones to a new entry that looks like the current FM1216ME MK3 entry unmodified. Andy, thanks a lot for participating in such stuff and I think your diagnosis is right. Just a small question in between, already late here and not trying to cover the whole scope. What ever should be the effect of Dmitri's trick one, changing beginning of UHF a little. We did that for midband and there was real broadcast and it improved one single channel there indeed. But here, it is plain theory. I honor the lab results they have, no problem anyway, but to change something for not at all existing broadcast does exactly nothing, except for awaiting it in the future. No problem with that change, but do I miss something? Now that you ask, maybe. It first depends on whether there is a station at 441 MHz that normally would have used the VHF-High filter and VCO, but now uses the UHF filter and VCO. Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. The high frequencies in the channel (~ 447 MHz) may have perhaps been in the roll-off of the VHF-High preselector filter. At the edges of filters, amplitude ripple and especially group delay variation - two aspects of filters that cause distortion - would have been at their worst, affecting the high frequencies of the channel (sound and color sub-carriers). (I assume PAL is VSB with the carrier towards the low end, similar to NTSC.) Now instead, the low frequencies of the channel (~ 440 MHz) may be in the roll-off of the UHF preselector filter. Thus the vestigal sideband and carrier could be affected most by ripple and group delay variation of the UHF filter. Either way, a channel at 440 MHz could face distortion by this tuner. It really depends on the preselector filter design. I also checked the MID and HIGH band oscillator spec's in the TUA6030 datasheet. Both of them can cover 440 MHz, but it looks like the MID band VCO may be preferred since it doesn't drift as badly as the HIGH band VCO. Since I don't know the component values used in the loop filters for the VCO's, I can't do any real analysis to see which VCO would be better at handling 440 MHz. I suspect the difference may not be significant anyway. Also, after hundreds of new tuners did appear, in the beginning not even known from where, I suggested to not allow a new tuner entry for all of them, only duplicate code, until they really need it and show off their difference. I would like to keep it especially for this one the same. ;) OK. Such subsumed under it have done nothing for Linux so far and have to face their faith :) And show off, if _not_ compatible. And not the other way round. Wait until people complain? :) Regards, Andy Dmitri, if we are talking about the same tuner and filters, we should try to get Secam D/K improvements into the original tuner entry. That NTSC hack stuff might go elsewhere I guess. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Am Freitag, den 08.05.2009, 22:01 -0400 schrieb Andy Walls: On Fri, 2009-05-08 at 02:08 +0200, hermann pitton wrote: Am Mittwoch, den 06.05.2009, 23:03 -0400 schrieb Andy Walls: For the change of UHF start I don't see any problem. If you're talking about the frequency for the bandswitch, I don't see a problem either in general. It may cause a problem for clones of the FM1216ME MK3 that don't have the same filter performance near the cutover, but use the same tuner defintion as the FM1216ME MK3 in tuner-types.c. It may be best to point any clones to a new entry that looks like the current FM1216ME MK3 entry unmodified. Andy, thanks a lot for participating in such stuff and I think your diagnosis is right. Just a small question in between, already late here and not trying to cover the whole scope. What ever should be the effect of Dmitri's trick one, changing beginning of UHF a little. We did that for midband and there was real broadcast and it improved one single channel there indeed. But here, it is plain theory. I honor the lab results they have, no problem anyway, but to change something for not at all existing broadcast does exactly nothing, except for awaiting it in the future. No problem with that change, but do I miss something? Now that you ask, maybe. It first depends on whether there is a station at 441 MHz that normally would have used the VHF-High filter and VCO, but now uses the UHF filter and VCO. Channel designations I dug out of ivtv-tune: S38 439.250 MHz (European cable) H18 439.250 MHz (SECAM France) 47 440.250 MHz (PAL China) 059 440.250 MHz (PAL Argentina) come close, but are unaffected by the change from 442 to 441 as the bandswitch cutover point. These channels fall right on top of the cutover, but are not affected by the proposed change in any meaningful way. The VHF-High filter and VCO would still be used. Dmitri's proposed change is a don't care unless the cutover point is changed to 440 MHz. Let's pretend that the proposed cutover point is 440 MHz. The high frequencies in the channel (~ 447 MHz) may have perhaps been in the roll-off of the VHF-High preselector filter. At the edges of filters, amplitude ripple and especially group delay variation - two aspects of filters that cause distortion - would have been at their worst, affecting the high frequencies of the channel (sound and color sub-carriers). (I assume PAL is VSB with the carrier towards the low end, similar to NTSC.) Now instead, the low frequencies of the channel (~ 440 MHz) may be in the roll-off of the UHF preselector filter. Thus the vestigal sideband and carrier could be affected most by ripple and group delay variation of the UHF filter. Either way, a channel at 440 MHz could face distortion by this tuner. It really depends on the preselector filter design. I also checked the MID and HIGH band oscillator spec's in the TUA6030 datasheet. Both of them can cover 440 MHz, but it looks like the MID band VCO may be preferred since it doesn't drift as badly as the HIGH band VCO. Since I don't know the component values used in the loop filters for the VCO's, I can't do any real analysis to see which VCO would be better at handling 440 MHz. I suspect the difference may not be significant anyway. Also, after hundreds of new tuners did appear, in the beginning not even known from where, I suggested to not allow a new tuner entry for all of them, only duplicate code, until they really need it and show off their difference. I would like to keep it especially for this one the same. ;) OK. Such subsumed under it have done nothing for Linux so far and have to face their faith :) And show off, if _not_ compatible. And not the other way round. Wait until people complain? :) Yes, of course. They had it for free and that is the trap. Again, don't try to start it the other way round. I still wait for Dmitri, to show really off what he has. And, unfortunately, I don't have anything to test on currently :) Cheers, Hermann Regards, Andy Dmitri, if we are talking about the same tuner and filters, we should try to get Secam D/K improvements into the original tuner entry. That NTSC hack stuff might go elsewhere I guess. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi Andy Sorry me delay. I discuss about it with our programmer. On Wed, 2009-04-29 at 20:12 +1000, Dmitri Belimov wrote: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? Dmitri, From my understanding the take over point is the signal strength level that you want the second stage (an IF demod chip like the TDA9887) to take over automatic gain control from the first stage (a tuner chip like the TUA6030). The objective is to set the TOP to achieve maximum gain in the first stage while avoiding clipping in the first stage. When the input signal level is smaller than the TOP, the first stage gain is at a maximum, and the second stage is performing automatic gain control internally. When the input signal level becomes greater than the TOP, the first stage gain needs to be reduced by an AGC, and the second stage gain remains constant. As I understand it, it would be best to set the first stage (TUA6030 or similar) to External AGC and set the take over point in the second stage (the TDA9887), if the pins between the chips are wired up properly inside the tuner. This should coordinate the AGC in both the first stage and second stage of the tuner, as the second stage will be providing the gain control to the first stage as needed when the signal reaches the TOP. http://www.comsec.com/usrp/microtune/NF_tutorial.pdf It allows you to achieve maximum gain in the first stage to minimize overall receiver noise figure, and avoid clipping the input signal in the first stage (TUA6030) with a proper TOP setting in the second stage (TDA9887). The TOP setting in the second stage needs to take into account IF SAW filter attenuation of course. Do the circuit board traces in the FM1216ME_MK3 support the TDA9887 controlling the gain of the first stage? (I've never opened an equivalent NTSC tuner assembly to take a look.) If not, then, if I understand things correctly, you need to set the first stage and second stage TOP settings so that they refer to about the same signal level before the IF SAW filter. You are right. My patch set AGC TOP level for RF signal in TUA6030. The TOP of TDA9887 control only IF signal. If set high level of AGC TOP for RF and RF signal is strengh, we can see white horizontal junk on TV image. For removing this junk need reduce AGC TOP RF for this channel. It can be only with Secam channels. With my best regards, Dmitry. I would think AGC TOP settings, for both stages of the tuner, are tuner-dependent and relatively constant once you figure out what they should be. Do you have a different understanding or insight? Regards, Andy If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-simple.c Tue Apr 21 09:44:38 2009 +1000 @@ -116,6 +116,7 @@ u32 frequency; u32 bandwidth; + signed int gain; }; /*
Re: [PATCH] FM1216ME_MK3 some changes
Am Mittwoch, den 06.05.2009, 23:03 -0400 schrieb Andy Walls: On Thu, 2009-05-07 at 02:01 +0200, hermann pitton wrote: Hi, Am Mittwoch, den 06.05.2009, 04:42 +1000 schrieb Dmitri Belimov: Hi Hermann Hi Dmitry, Am Mittwoch, den 29.04.2009, 20:12 +1000 schrieb Dmitri Belimov: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. From the tda9887_3 datasheet. 8.2 Tuner AGC and VIF-AGC This block adapts the voltages, generated at the VIF-AGC and SIF-AGC detectors, to the internal signal processing at the VIF and SIF amplifiers and performs the tuner AGC control current generation. The onset of the tuner AGC control current generation can be set either via the I2C-bus (see Table 13) or optionally by a potentiometer at pin TOP (in case that the I2C-bus information cannot be stored). The presence of a potentiometer is automatically detected and the I2C-bus setting is disabled. Reads for me that on some tuners, like NTSC only, the tuner AGC is fix. To change it per channel is not needed at all. I've started looking at the photographs of the tuner that Hermann sent. Looking at the TDA9887 v4 datasheet, I can see how the TOP related pins are (not) wired to the 1st stage oscillator/mixer chip. Pin 9 (TOP) is supposed to either be tied to ground through a resistor or left open. This design has the pin connected to a white SMT capacitor(?) which I will assume the TDA9887 will see as an open circuit. This means the TOP in the TDA9887 should be programmable via I2C. Pin 14 (TAGC) looks like it is unconnected. This means the TDA9887 TAGC output is not actively taking over gain control of the 1st stage RF mixer/oscialltor chip. So this answers a question I had posed to Dmitry: Is this tuner wired up so the TDA9887 can adjust the gain of the 1st stage? That answer is no. That means, that in this tuner, the AGC's in the two chips are operating independently and need their TOP's set in a coordinated manner that takes into account the IF filter losses and the modulation. Since the FM1216ME data sheet makes recommendations for TOP values for both chips: RF mixer/osc chip: 109 dBμV Recommended for negative modulation 106 dBμV Recommended for positive modulation It is recommended to set the TOP at 109 dBμV for PAL B/G, D/K, I For system L/L’, it is recommended to set the TOP at 106 dBμV. For FM radio, it is also recommended to set the TOP to 109 dBμV IF demod chip: C4-C0 = 1 [0 dB, 17 mV (RMS) according to the TDA9887 datasheet] for all FM modes and B/G D/K I L and L' systems it is probably best to use those. I can try and look at the IF filter from the picutres Hermann sent, but I'm not sure I'd be able to figure out the loss and come up with better TOP setting recommendations than the manufacturer's datasheet. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: what about to create a new FM1216ME MK3 tuner type for testing? Yes. Can you do it? I start add MK5 tuner. Yes, if really needed, but I see no hint anywhere that it should be useful to have TOP settings per channel for user applications and will stay away from it. Right now, I don't think the tuner-simple.c code adjusts the TOP value based on the video system or FM mode. It probably should. I don't think a user control will be very useful. For the change of UHF start I don't see any problem. If you're talking about the frequency for the bandswitch, I don't see a problem either in general. It may cause a problem for clones of the FM1216ME MK3 that don't
Re: [PATCH] FM1216ME_MK3 some changes
Hi Hermann Hi Dmitry, Am Mittwoch, den 29.04.2009, 20:12 +1000 schrieb Dmitri Belimov: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: what about to create a new FM1216ME MK3 tuner type for testing? Yes. Can you do it? I start add MK5 tuner. It is for sure maybe the best tuner around and likely is also the best for tweaks in countries with huge lands to cover, not restricted to Russia, Australia or Canada. All our tuners tested with all standarts SECAM, PAL, NTSC because big country, some people want receive TV from Korea and Japan. For tests, our hardware engineer has hardware TV generator. With my best regards, Dmitry. I would like to have also Hans' opinion on it, since he did some final steps to get them right. Cheers, Hermann 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-simple.c Tue Apr 21 09:44:38 2009 +1000 @@ -116,6 +116,7 @@ u32 frequency; u32 bandwidth; + signed int gain; }; /* -- */ @@ -495,15 +507,57 @@ (should be 4)\n, rc); break; } + case TUNER_PHILIPS_FM1216ME_MK3: + { + buffer[2] = 0xDE; /* T2 = 0, T1 = 1 and T0 = 1 */ + switch (priv-gain) { + case 0: + /* TOP = External AGC, ATC = OFF */ + buffer[3] = 0x60; + break; + case 1: + /* TOP = 118 dB, ATC = OFF */ + buffer[3] = 0x00; + break; + case 2: + /* TOP = 115 dB, ATC = OFF */ + buffer[3] = 0x10; + break; + case 3: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + case 4: + /* TOP = 109 dB, ATC = OFF */ + buffer[3] = 0x30; + break; + case 5: + /* TOP = 106 dB, ATC = OFF */ + buffer[3] = 0x40; + break; + case 6: + /* TOP = 103 dB, ATC = OFF */ + buffer[3] = 0x50; + break; + default: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + } + + tuner_dbg(tv 0x%02x 0x%02x 0x%02x 0x%02x\n, + buffer[0], buffer[1], buffer[2], buffer[3]); + + rc = tuner_i2c_xfer_send(priv-i2c_props, buffer, 4); + if (4 != rc) + tuner_warn(i2c i/o error: rc == %d + (should be 4)\n, rc); + + break; } - + } return 0; } diff -r 43dbc8ebb5a2
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Mittwoch, den 06.05.2009, 04:42 +1000 schrieb Dmitri Belimov: Hi Hermann Hi Dmitry, Am Mittwoch, den 29.04.2009, 20:12 +1000 schrieb Dmitri Belimov: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. From the tda9887_3 datasheet. 8.2 Tuner AGC and VIF-AGC This block adapts the voltages, generated at the VIF-AGC and SIF-AGC detectors, to the internal signal processing at the VIF and SIF amplifiers and performs the tuner AGC control current generation. The onset of the tuner AGC control current generation can be set either via the I2C-bus (see Table 13) or optionally by a potentiometer at pin TOP (in case that the I2C-bus information cannot be stored). The presence of a potentiometer is automatically detected and the I2C-bus setting is disabled. Reads for me that on some tuners, like NTSC only, the tuner AGC is fix. To change it per channel is not needed at all. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: what about to create a new FM1216ME MK3 tuner type for testing? Yes. Can you do it? I start add MK5 tuner. Yes, if really needed, but I see no hint anywhere that it should be useful to have TOP settings per channel for user applications and will stay away from it. For the change of UHF start I don't see any problem. SECAM_DK is never mentioned explicitly in the old datasheet we have at ivtvdriver.org. Maybe we miss something here, but you try to change over the full range and simply 109 dB is recommended for negative demodulation. (page 13 for recommended TOP settings) http://dl.ivtvdriver.org/datasheets/tuners Maybe we are really talking about different tuner versions/revisions. Can you have a look at the tuner pictures I did send? It is for sure maybe the best tuner around and likely is also the best for tweaks in countries with huge lands to cover, not restricted to Russia, Australia or Canada. All our tuners tested with all standarts SECAM, PAL, NTSC because big country, some people want receive TV from Korea and Japan. For tests, our hardware engineer has hardware TV generator. That confuses me a lot :) You mean you test that tuner with its PAL/SECAM SAW filters on NTSC_KR and NTSC_JP ? With my best regards, Dmitry. I would like to have also Hans' opinion on it, since he did some final steps to get them right. Andy might have a look too. Please wait a little for more opinions. Cheers, Hermann 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.cTue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-simple.cTue Apr 21 09:44:38 2009 +1000 @@ -116,6 +116,7 @@ u32 frequency; u32 bandwidth; + signed int gain; }; /* -- */ @@ -495,15 +507,57 @@ (should be 4)\n, rc); break; } + case TUNER_PHILIPS_FM1216ME_MK3: + { + buffer[2] = 0xDE; /* T2 = 0, T1 = 1 and T0 = 1 */ +
Re: [PATCH] FM1216ME_MK3 some changes
On Thu, 2009-05-07 at 02:01 +0200, hermann pitton wrote: Hi, Am Mittwoch, den 06.05.2009, 04:42 +1000 schrieb Dmitri Belimov: Hi Hermann Hi Dmitry, Am Mittwoch, den 29.04.2009, 20:12 +1000 schrieb Dmitri Belimov: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. From the tda9887_3 datasheet. 8.2 Tuner AGC and VIF-AGC This block adapts the voltages, generated at the VIF-AGC and SIF-AGC detectors, to the internal signal processing at the VIF and SIF amplifiers and performs the tuner AGC control current generation. The onset of the tuner AGC control current generation can be set either via the I2C-bus (see Table 13) or optionally by a potentiometer at pin TOP (in case that the I2C-bus information cannot be stored). The presence of a potentiometer is automatically detected and the I2C-bus setting is disabled. Reads for me that on some tuners, like NTSC only, the tuner AGC is fix. To change it per channel is not needed at all. I've started looking at the photographs of the tuner that Hermann sent. Looking at the TDA9887 v4 datasheet, I can see how the TOP related pins are (not) wired to the 1st stage oscillator/mixer chip. Pin 9 (TOP) is supposed to either be tied to ground through a resistor or left open. This design has the pin connected to a white SMT capacitor(?) which I will assume the TDA9887 will see as an open circuit. This means the TOP in the TDA9887 should be programmable via I2C. Pin 14 (TAGC) looks like it is unconnected. This means the TDA9887 TAGC output is not actively taking over gain control of the 1st stage RF mixer/oscialltor chip. So this answers a question I had posed to Dmitry: Is this tuner wired up so the TDA9887 can adjust the gain of the 1st stage? That answer is no. That means, that in this tuner, the AGC's in the two chips are operating independently and need their TOP's set in a coordinated manner that takes into account the IF filter losses and the modulation. Since the FM1216ME data sheet makes recommendations for TOP values for both chips: RF mixer/osc chip: 109 dBμV Recommended for negative modulation 106 dBμV Recommended for positive modulation It is recommended to set the TOP at 109 dBμV for PAL B/G, D/K, I For system L/L’, it is recommended to set the TOP at 106 dBμV. For FM radio, it is also recommended to set the TOP to 109 dBμV IF demod chip: C4-C0 = 1 [0 dB, 17 mV (RMS) according to the TDA9887 datasheet] for all FM modes and B/G D/K I L and L' systems it is probably best to use those. I can try and look at the IF filter from the picutres Hermann sent, but I'm not sure I'd be able to figure out the loss and come up with better TOP setting recommendations than the manufacturer's datasheet. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: what about to create a new FM1216ME MK3 tuner type for testing? Yes. Can you do it? I start add MK5 tuner. Yes, if really needed, but I see no hint anywhere that it should be useful to have TOP settings per channel for user applications and will stay away from it. Right now, I don't think the tuner-simple.c code adjusts the TOP value based on the video system or FM mode. It probably should. I don't think a user control will be very useful. For the change of UHF start I don't see any problem. If you're talking about the frequency for the bandswitch, I don't see a problem either in general. It may cause a problem for clones of the FM1216ME MK3 that don't have the same filter performance near the cutover, but use the same tuner defintion as the FM1216ME MK3 in tuner-types.c. It may be best to point any clones to a new entry that looks like the
Re: [PATCH] FM1216ME_MK3 some changes
Am Samstag, den 02.05.2009, 04:05 +0200 schrieb hermann pitton: Am Freitag, den 01.05.2009, 20:52 -0400 schrieb Andy Walls: On Sat, 2009-05-02 at 01:55 +0200, hermann pitton wrote: Guys, [snip] Do the circuit board traces in the FM1216ME_MK3 support the TDA9887 controlling the gain of the first stage? (I've never opened an equivalent NTSC tuner assembly to take a look.) equivalent NTSC tuners _do not_ exist at all. I don't forget all the time we spend to find out that some of them are Intercarrier only! Also, the tda988x stuff is underneath the tuner PCB. I cut one off for those interested in line tracing ... still on the to do list ;) Thanks. I hope it doesn't cost you too many wasted Euros... Seven years later it is just trash and i hope to provide enough for everyone ;) Andy, did cut out my old one after lunch and have already acceptable picture quality under daylight conditions for the not yet seen backside. For the surface with the tin compartments I must either find my old pictures sent to Gunther again or make some new with setting some lights to avoid harsh shadows. If I'm done, and nobody has objections, I plan to send to you, Mauro, Hans and Hartmut off list. Also to Dmitri, maybe he has a newer revision, and to CityK and Gunther for documentation. Cheers, Hermann Without port2=0 you don't get any SECAM-L into the sound trap. It needs amplification from minus 40 dB AM for the first sound carrier, and then of course you prefer the second with NICAM. If not, then, if I understand things correctly, you need to set the first stage and second stage TOP settings so that they refer to about the same signal level before the IF SAW filter. I would think AGC TOP settings, for both stages of the tuner, are tuner-dependent and relatively constant once you figure out what they should be. Do you have a different understanding or insight? Regards, Andy Hartmut once offered to make contacts with colleagues at Philips Hamburg for such tuners and related tda9887 stuff. Unfortunately he is not active on the lists currently. If I see, how easily someone can get a patch to Andrew and disables all other SECAM stuff, again from Russia, I'm not convinced on anything from there. I seriously doubt that those tuners are meant for fumbling on TOP RF/IF settings from user space. The proper TOP settings should be a function of the input signal TV standard (positive modulation vs. negative modulation vs. FM radio) and the components in the tuner assembly (the IF filter for sure). Setting TOP settings from user space only makes sense for laboratory experimentation. Linux is kind of funny in the way it breaks up the integrated assembly of an analog tuner to be handled by separate drivers: preselector and mixer/osc chip by tuner-simple; IF filter and IF demod by another module like tda9887. Really the two chips should be set with coordinated settings, taking into account the TV standard the tuner is being set to demodulate and the properties of the IF filter and external components to the IF demodulator. The linux analog tuner drivers mostly appear to make some safe, default settings, and don't attempt to get the best performance. I'm assuming that is due to lack of data on the tuners and the tuners working well enough. I'm glad to see Dmitri working at trying to improve analog performance. I wish I had the proper equpment to experiment wih NTSC tiners. If I had a lab with a decent signal generator, oscilliscope, etc. I'd probably have lots of fun playing around with these tuners. :) Regards, Andy -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
On Sat, 2009-05-02 at 01:55 +0200, hermann pitton wrote: Guys, [snip] Do the circuit board traces in the FM1216ME_MK3 support the TDA9887 controlling the gain of the first stage? (I've never opened an equivalent NTSC tuner assembly to take a look.) equivalent NTSC tuners _do not_ exist at all. I don't forget all the time we spend to find out that some of them are Intercarrier only! Also, the tda988x stuff is underneath the tuner PCB. I cut one off for those interested in line tracing ... still on the to do list ;) Thanks. I hope it doesn't cost you too many wasted Euros... Without port2=0 you don't get any SECAM-L into the sound trap. It needs amplification from minus 40 dB AM for the first sound carrier, and then of course you prefer the second with NICAM. If not, then, if I understand things correctly, you need to set the first stage and second stage TOP settings so that they refer to about the same signal level before the IF SAW filter. I would think AGC TOP settings, for both stages of the tuner, are tuner-dependent and relatively constant once you figure out what they should be. Do you have a different understanding or insight? Regards, Andy Hartmut once offered to make contacts with colleagues at Philips Hamburg for such tuners and related tda9887 stuff. Unfortunately he is not active on the lists currently. If I see, how easily someone can get a patch to Andrew and disables all other SECAM stuff, again from Russia, I'm not convinced on anything from there. I seriously doubt that those tuners are meant for fumbling on TOP RF/IF settings from user space. The proper TOP settings should be a function of the input signal TV standard (positive modulation vs. negative modulation vs. FM radio) and the components in the tuner assembly (the IF filter for sure). Setting TOP settings from user space only makes sense for laboratory experimentation. Linux is kind of funny in the way it breaks up the integrated assembly of an analog tuner to be handled by separate drivers: preselector and mixer/osc chip by tuner-simple; IF filter and IF demod by another module like tda9887. Really the two chips should be set with coordinated settings, taking into account the TV standard the tuner is being set to demodulate and the properties of the IF filter and external components to the IF demodulator. The linux analog tuner drivers mostly appear to make some safe, default settings, and don't attempt to get the best performance. I'm assuming that is due to lack of data on the tuners and the tuners working well enough. I'm glad to see Dmitri working at trying to improve analog performance. I wish I had the proper equpment to experiment wih NTSC tiners. If I had a lab with a decent signal generator, oscilliscope, etc. I'd probably have lots of fun playing around with these tuners. :) Regards, Andy Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Am Freitag, den 01.05.2009, 20:52 -0400 schrieb Andy Walls: On Sat, 2009-05-02 at 01:55 +0200, hermann pitton wrote: Guys, [snip] Do the circuit board traces in the FM1216ME_MK3 support the TDA9887 controlling the gain of the first stage? (I've never opened an equivalent NTSC tuner assembly to take a look.) equivalent NTSC tuners _do not_ exist at all. I don't forget all the time we spend to find out that some of them are Intercarrier only! Also, the tda988x stuff is underneath the tuner PCB. I cut one off for those interested in line tracing ... still on the to do list ;) Thanks. I hope it doesn't cost you too many wasted Euros... Seven years later it is just trash and i hope to provide enough for everyone ;) Cheers, Hermann Without port2=0 you don't get any SECAM-L into the sound trap. It needs amplification from minus 40 dB AM for the first sound carrier, and then of course you prefer the second with NICAM. If not, then, if I understand things correctly, you need to set the first stage and second stage TOP settings so that they refer to about the same signal level before the IF SAW filter. I would think AGC TOP settings, for both stages of the tuner, are tuner-dependent and relatively constant once you figure out what they should be. Do you have a different understanding or insight? Regards, Andy Hartmut once offered to make contacts with colleagues at Philips Hamburg for such tuners and related tda9887 stuff. Unfortunately he is not active on the lists currently. If I see, how easily someone can get a patch to Andrew and disables all other SECAM stuff, again from Russia, I'm not convinced on anything from there. I seriously doubt that those tuners are meant for fumbling on TOP RF/IF settings from user space. The proper TOP settings should be a function of the input signal TV standard (positive modulation vs. negative modulation vs. FM radio) and the components in the tuner assembly (the IF filter for sure). Setting TOP settings from user space only makes sense for laboratory experimentation. Linux is kind of funny in the way it breaks up the integrated assembly of an analog tuner to be handled by separate drivers: preselector and mixer/osc chip by tuner-simple; IF filter and IF demod by another module like tda9887. Really the two chips should be set with coordinated settings, taking into account the TV standard the tuner is being set to demodulate and the properties of the IF filter and external components to the IF demodulator. The linux analog tuner drivers mostly appear to make some safe, default settings, and don't attempt to get the best performance. I'm assuming that is due to lack of data on the tuners and the tuners working well enough. I'm glad to see Dmitri working at trying to improve analog performance. I wish I had the proper equpment to experiment wih NTSC tiners. If I had a lab with a decent signal generator, oscilliscope, etc. I'd probably have lots of fun playing around with these tuners. :) Regards, Andy Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Am Mittwoch, den 29.04.2009, 21:14 -0400 schrieb Andy Walls: On Wed, 2009-04-29 at 20:12 +1000, Dmitri Belimov wrote: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? Dmitri, From my understanding the take over point is the signal strength level that you want the second stage (an IF demod chip like the TDA9887) to take over automatic gain control from the first stage (a tuner chip like the TUA6030). The objective is to set the TOP to achieve maximum gain in the first stage while avoiding clipping in the first stage. When the input signal level is smaller than the TOP, the first stage gain is at a maximum, and the second stage is performing automatic gain control internally. When the input signal level becomes greater than the TOP, the first stage gain needs to be reduced by an AGC, and the second stage gain remains constant. As I understand it, it would be best to set the first stage (TUA6030 or similar) to External AGC and set the take over point in the second stage (the TDA9887), if the pins between the chips are wired up properly inside the tuner. This should coordinate the AGC in both the first stage and second stage of the tuner, as the second stage will be providing the gain control to the first stage as needed when the signal reaches the TOP. http://www.comsec.com/usrp/microtune/NF_tutorial.pdf It allows you to achieve maximum gain in the first stage to minimize overall receiver noise figure, and avoid clipping the input signal in the first stage (TUA6030) with a proper TOP setting in the second stage (TDA9887). The TOP setting in the second stage needs to take into account IF SAW filter attenuation of course. Do the circuit board traces in the FM1216ME_MK3 support the TDA9887 controlling the gain of the first stage? (I've never opened an equivalent NTSC tuner assembly to take a look.) equivalent NTSC tuners _do not_ exist at all. I don't forget all the time we spend to find out that some of them are Intercarrier only! Also, the tda988x stuff is underneath the tuner PCB. I cut one off for those interested in line tracing ... Without port2=0 you don't get any SECAM-L into the sound trap. It needs amplification from minus 40 dB AM for the first sound carrier, and then of course you prefer the second with NICAM. If not, then, if I understand things correctly, you need to set the first stage and second stage TOP settings so that they refer to about the same signal level before the IF SAW filter. I would think AGC TOP settings, for both stages of the tuner, are tuner-dependent and relatively constant once you figure out what they should be. Do you have a different understanding or insight? Regards, Andy Since I have some m$ system again after 9 years, not used within the last three months, I would prefer to get it demonstrated there at first. I leave on the first BSOD. Cheers, Hermann If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.c Tue Jan 27
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-simple.c Tue Apr 21 09:44:38 2009 +1000 @@ -116,6 +116,7 @@ u32 frequency; u32 bandwidth; + signed int gain; }; /* -- */ @@ -495,15 +507,57 @@ (should be 4)\n, rc); break; } + case TUNER_PHILIPS_FM1216ME_MK3: + { + buffer[2] = 0xDE; /* T2 = 0, T1 = 1 and T0 = 1 */ + switch (priv-gain) { + case 0: + /* TOP = External AGC, ATC = OFF */ + buffer[3] = 0x60; + break; + case 1: + /* TOP = 118 dB, ATC = OFF */ + buffer[3] = 0x00; + break; + case 2: + /* TOP = 115 dB, ATC = OFF */ + buffer[3] = 0x10; + break; + case 3: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + case 4: + /* TOP = 109 dB, ATC = OFF */ + buffer[3] = 0x30; + break; + case 5: + /* TOP = 106 dB, ATC = OFF */ + buffer[3] = 0x40; + break; + case 6: + /* TOP = 103 dB, ATC = OFF */ + buffer[3] = 0x50; + break; + default: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + } + + tuner_dbg(tv 0x%02x 0x%02x 0x%02x 0x%02x\n, + buffer[0], buffer[1], buffer[2], buffer[3]); + + rc = tuner_i2c_xfer_send(priv-i2c_props, buffer, 4); + if (4 != rc) + tuner_warn(i2c i/o error: rc == %d + (should be 4)\n, rc); + + break; } - + } return 0; } diff -r 43dbc8ebb5a2 linux/drivers/media/video/saa7134/saa7134-cards.c --- a/linux/drivers/media/video/saa7134/saa7134-cards.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/video/saa7134/saa7134-cards.c Tue Apr 21 09:44:38 2009 +1000 @@ -6506,6 +6806,20 @@ saa_call_all(dev, tuner, s_config, tea5767_cfg); break; } + case SAA7134_BOARD_BEHOLD_M6_EXTRA: + { + struct v4l2_queryctrl *ctl; + struct saa7134_fh *fh; + struct file *fl; + + ctl-id = V4L2_CID_GAIN; + if (saa7134_queryctrl(fl, fh, ctl) == 0) {/* BUG here */ + /* enable this control */ + ctl-flags = ~(V4L2_CTRL_FLAG_DISABLED); +
Re: [PATCH] FM1216ME_MK3 some changes
Hi Dmitry, Am Mittwoch, den 29.04.2009, 20:12 +1000 schrieb Dmitri Belimov: Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. This is my idea for RFC about TOP AGC: what about to create a new FM1216ME MK3 tuner type for testing? It is for sure maybe the best tuner around and likely is also the best for tweaks in countries with huge lands to cover, not restricted to Russia, Australia or Canada. I would like to have also Hans' opinion on it, since he did some final steps to get them right. Cheers, Hermann 1. Add gain variable to tuner structure. 2. Add V4L2_CID_GAIN control to saa7134 and disable this control. 3. Add workaround to simple_post_tune function for write sensitivity level to the tuner. 4. Enable V4L2_CID_GAIN control when module load if card is right. My expirience not so good, step 4 segfault the kernel. How to we can make it? Our windows end-user programm control the sensitivity of each TV channel and change when channel changed. What you think about it?? If TV card is not touch V4L2_CTRL_FLAG_DISABLED in this control. The programm can't change AGC TOP. And write default value to AGC TOP like now. diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-simple.c --- a/linux/drivers/media/common/tuners/tuner-simple.cTue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-simple.cTue Apr 21 09:44:38 2009 +1000 @@ -116,6 +116,7 @@ u32 frequency; u32 bandwidth; + signed int gain; }; /* -- */ @@ -495,15 +507,57 @@ (should be 4)\n, rc); break; } + case TUNER_PHILIPS_FM1216ME_MK3: + { + buffer[2] = 0xDE; /* T2 = 0, T1 = 1 and T0 = 1 */ + switch (priv-gain) { + case 0: + /* TOP = External AGC, ATC = OFF */ + buffer[3] = 0x60; + break; + case 1: + /* TOP = 118 dB, ATC = OFF */ + buffer[3] = 0x00; + break; + case 2: + /* TOP = 115 dB, ATC = OFF */ + buffer[3] = 0x10; + break; + case 3: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + case 4: + /* TOP = 109 dB, ATC = OFF */ + buffer[3] = 0x30; + break; + case 5: + /* TOP = 106 dB, ATC = OFF */ + buffer[3] = 0x40; + break; + case 6: + /* TOP = 103 dB, ATC = OFF */ + buffer[3] = 0x50; + break; + default: + /* TOP = 112 dB, ATC = OFF */ + buffer[3] = 0x20; + break; + } + + tuner_dbg(tv 0x%02x 0x%02x 0x%02x 0x%02x\n, + buffer[0], buffer[1], buffer[2], buffer[3]); + + rc = tuner_i2c_xfer_send(priv-i2c_props, buffer, 4); + if (4 != rc) + tuner_warn(i2c i/o error: rc == %d +(should be 4)\n, rc); + + break; } - + } return 0; } diff -r 43dbc8ebb5a2 linux/drivers/media/video/saa7134/saa7134-cards.c --- a/linux/drivers/media/video/saa7134/saa7134-cards.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/video/saa7134/saa7134-cards.c Tue Apr 21 09:44:38 2009 +1000 @@ -6506,6 +6806,20 @@ saa_call_all(dev, tuner,
Re: [PATCH] FM1216ME_MK3 some changes
On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. Cheers, Mauro. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy Cheers, Mauro -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. With my best regards, Dmitry. Cheers, Mauro. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Dienstag, den 28.04.2009, 19:59 +1000 schrieb Dmitri Belimov: On Tue, 28 Apr 2009 15:18:32 -0300 Mauro Carvalho Chehab mche...@infradead.org wrote: On Mon, 27 Apr 2009 19:29:05 +1000 Dmitri Belimov d.beli...@gmail.com wrote: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. Dmitri, I'll mark those patches as RFC at patchwork until the end of those discussions. After that, please send it again into a new thread. You mark patch with TOP AGC not this. I think need discuss about FM1216ME_MK3 because I'll have a big patch for support control TOP AGC (sensitivity) of this tuner. It can be bad for compatible tuners. hmm, in Europe, that TOP AGC did not ever made much difference and it is an insmod option since ever. I can't tell for Sibiria and initially that tuner had no SECAM-DK support officially at all. There are no good/much_better tuners for FTA at all and never have been ;) Some examples, user success reports, to make it more easily to understand? I think it can only change some _very little_ under already worst conditions. Cheers, Hermann With my best regards, Dmitry. Cheers, Mauro. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise
Re: [PATCH] FM1216ME_MK3 some changes
Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Montag, den 27.04.2009, 19:29 +1000 schrieb Dmitri Belimov: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com if it does help anything for having all above 441 MHz in UHF ranges. Acked-by: Hermann Pitton hermann-pit...@arcor.de There is simply no known broadcast in that gap on other freq. tables. Can you please enlighten us who/what broadcasts there? With my best regards, Dmitry. Hi Dmitri, Cheers, Hermann Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Am Dienstag, den 28.04.2009, 02:11 +0200 schrieb hermann pitton: Hi, Am Montag, den 27.04.2009, 19:29 +1000 schrieb Dmitri Belimov: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com if it does help anything for having all above 441 MHz in UHF ranges. for the record: from 441 MHz on ... Acked-by: Hermann Pitton hermann-pit...@arcor.de There is simply no known broadcast in that gap on other freq. tables. Can you please enlighten us who/what broadcasts there? With my best regards, Dmitry. Hi Dmitri, Cheers, Hermann Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Montag, den 27.04.2009, 20:42 +1000 schrieb Dmitri Belimov: Hi, Am Montag, den 27.04.2009, 19:29 +1000 schrieb Dmitri Belimov: Hi All Step by step. This is patch for change only range of FM1216ME_MK3. Slow tunning is not a big problem. diff -r b40d628f830d linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Fri Apr 24 01:46:41 2009 -0300 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 28 03:35:42 2009 +1000 @@ -558,8 +558,8 @@ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 441.00 /*MHz*/, 0x8e, 0x02, }, + { 16 * 864.00, 0x8e, 0x04, }, }; static struct tuner_params tuner_fm1216me_mk3_params[] = { Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com if it does help anything for having all above 441 MHz in UHF ranges. Acked-by: Hermann Pitton hermann-pit...@arcor.de There is simply no known broadcast in that gap on other freq. tables. Can you please enlighten us who/what broadcasts there? I don't know. This trick for solve problem with low sensitivity the tuners on end of band. We get it from measuments. I believe that, since seen before on real broadcast in the VHF low/high gap. But if there is no real broadcast at all, what it is all about for now? You can find this problem in theory on almost all new tuners, on every TCL, TENA, YMEC and what the hell else we have melt down to be in fact only _very few_ new ones. Get out of the lab, until it is a real problem, but keep in mind. Cheers, Hermann With my best regards, Dmitry. Hi Dmitri, Cheers, Hermann With my best regards, Dmitry. Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in
Re: [PATCH] FM1216ME_MK3 some changes
Hi Andy Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) ;-) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. We can make fake FM1216ME_MK3 compatible tuner for clones. And move clones to it. After testing each clone with real compatibility with FM1216ME_MK3 we can move to real tuner. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. Ok. I'll discuss about it with our system programmer. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Yes, you are right. This solutions used for hybrid tuners too. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. Regards, Andy With my best regards, Dmitry. -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Sonntag, den 26.04.2009, 23:58 +1000 schrieb Dmitri Belimov: Hi Andy Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) ;-) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3,TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. We can make fake FM1216ME_MK3 compatible tuner for clones. And move clones to it. After testing each clone with real compatibility with FM1216ME_MK3 we can move to real tuner. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. Some confusion probably started here. You were talking about one more different tuner ... I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. Ok. I'll discuss about it with our system programmer. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Yes, you are right. This solutions used for hybrid tuners too. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. Regards, Andy With my best regards, Dmitry. -- Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. This concerns me slightly as it does not match the datasheet (hence the design objectives) of the FM1236ME_MK3. Yes, I know. How are you measuring sensitivity? Do you know if it really is the middle-band preselector filter (and PLL and Mixer) or is it a problem with the input signal? How do you know it is not manufacturing variations in the preselector filters with the particular tuner assembly you are testing? Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. Also, as an alternative to using a different frequency for the bandswitch, have you considered setting the Auxillary Byte in the tuner chip (Infineon TUA6030?) to use external AGC and experimented with changing the tuner AGC take-over point (TOP) in the TDA9887? By maximizing the gain in the tuner chip, but avoiding clipping, with the proper TOP setting, you minimize the contributions by the rest of the receive chain to the overall receiver Noise Figure: http://en.wikipedia.org/wiki/Friis_formulas_for_noise This may be a way to improve receiver sensitivity that does not conflict with the data sheet specification. 2. Set correct highest freq of the higher band. :) This bothers me too; all the tuners in tuner-types.c have it set too high (999.0 MHz). I think I rememeber at time when all the tuner_range definitions had a real value there. It would be nice to have a real value there for all the tuners. The function tuner-simple.c:simple_config_lookup() would then prevent attempts to tune to an unsupported frequnecy. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. But in digital TV mode noise from PLL decreased BER. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. With my best regards, Dmitry. Regards, Andy diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 21 09:44:38 2009 +1000 @@ -557,9 +557,9 @@ /* TUNER_PHILIPS_FM1216ME_MK3 - Philips PAL */ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { - { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 158.00 /*MHz*/, 0xc6, 0x01, }, + { 16 * 441.00 /*MHz*/, 0xc6, 0x02, }, + { 16 * 864.00, 0xc6, 0x04, }, }; Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. -- video4linux-list mailing list Unsubscribe mailto:video4linux-list-requ...@redhat.com?subject=unsubscribe https://www.redhat.com/mailman/listinfo/video4linux-list -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi, Am Donnerstag, den 23.04.2009, 21:43 -0400 schrieb Andy Walls: Hi Dmitri, Thank you for you responses. Just a few more comments... On Thu, 2009-04-23 at 20:36 +1000, Dmitri Belimov wrote: Hi Andy Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. Several years a go your customers write some messages about bad quality of TV if frequency of TV is the end of band. It can be low band or middle. Our hardware engeneer make some tests with hardware TV generator and our TV tuners. If we set default frequency range for low and middle band, quality of TV signal on 159MHz and 442 MHz is bad. When we make our changes with moving end of bands the quality of TV much better. And our system programmer for OS Windows use changed bands for drivers. Customers be happy. OK. A properly run experiment wins over theory every time. :) You can test it if in your placement available TV programm on 159MHz or 442MHz. This trick can be usefull for other tuners. If you look at tveeprom.c, a number of other tuners are using that tuner definition: $ grep FM1216ME_MK3 tveeprom.c { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216 ME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S001D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, LG S701D MK3}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FQ1216LME MK3}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MFPE05 2}, { TUNER_PHILIPS_FM1216ME_MK3, TCL MPE05-2}, { TUNER_PHILIPS_FM1216ME_MK3, Philips FM1216ME MK5}, If your change makes things bad for the other tuners, we'll probably have to create an alternate entry for the other tuners instead of using the FM1216ME_MK3 defintion. I suspect most of them are clones of the FM1216ME MK3 however, so it probably won't matter. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. We discuss with our windows system programmer about it. He sad that in analog TV mode noise from PLL don't give any problem. I would be concerned about phase noise affecting the colors or any FM sound carriers. If the noise isn't noticably affecting colors to the human eye (do color bars look OK?), or sound to the human ear, then OK. But in digital TV mode noise from PLL decreased BER. I thought the FM1216ME MK3 was an analog only tuner. I guess I don't know DVB-T or cable in Europe well enough. that is for sure, analog only. Dmitry is preparing something for the MK5 too and the FMD1216ME/I MK3 hybrid. Hmm, I wonder if Hans and Jarod do have something to improve for the MK4 too and the others. As said, I don't care for changes within the freq. gap and accept everything working better there ;) Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Yes. You are right. We can swith CP off in FM mode. OK. Thank you. With my best regards, Dmitry. Regards, Andy Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] FM1216ME_MK3 some changes
Hi Andy and Dmitry, Am Mittwoch, den 22.04.2009, 22:08 -0400 schrieb Andy Walls: Dmitri, On Wed, 2009-04-22 at 17:48 +1000, Dmitri Belimov wrote: Hi All 1. Change middle band. In the end of the middle band the sensitivity of receiver not good. If we switch to higher band, sensitivity more better. Hardware trick. first of all, Dmitry, you need to send all your patches, also the prior ones, to linux-me...@vger.kernel.org. Only there patchwork will parse for them and you might eventually find them here. http://patchwork.kernel.org/project/linux-media/list http://patchwork.kernel.org/project/linux-media/list/?state=* You seem to have missed that completely and thanks to Andy to start to comment on this one, even only on video4linux previously. This concerns me slightly as it does not match the datasheet (hence the design objectives) of the FM1236ME_MK3. How are you measuring sensitivity? Do you know if it really is the middle-band preselector filter (and PLL and Mixer) or is it a problem with the input signal? How do you know it is not manufacturing variations in the preselector filters with the particular tuner assembly you are testing? I did not look into the datasheet again, we did not have it at all for a very long time, but we had a similar case already and that time we made the decision like it can be found in this thread. No complaints ever until today. http://marc.info/?l=linux-videom=112639247330257w=2 Did not look up all frequency tables again, but if it is the same again, I would say the risk of doing something wrong is close to zero. For the rest I never noticed any difference and we have it on lots of other tuners like that. And, that one still means a lot of _different_ tuners and different manufacturers, filters definitely differ! ... Also, as an alternative to using a different frequency for the bandswitch, have you considered setting the Auxillary Byte in the tuner chip (Infineon TUA6030?) to use external AGC and experimented with changing the tuner AGC take-over point (TOP) in the TDA9887? By maximizing the gain in the tuner chip, but avoiding clipping, with the proper TOP setting, you minimize the contributions by the rest of the receive chain to the overall receiver Noise Figure: http://en.wikipedia.org/wiki/Friis_formulas_for_noise This may be a way to improve receiver sensitivity that does not conflict with the data sheet specification. 2. Set correct highest freq of the higher band. :) This bothers me too; all the tuners in tuner-types.c have it set too high (999.0 MHz). I think I rememeber at time when all the tuner_range definitions had a real value there. It would be nice to have a real value there for all the tuners. The function tuner-simple.c:simple_config_lookup() would then prevent attempts to tune to an unsupported frequnecy. 3. Set charge pump bit This will improve the time to initially tune to a frequency, but will likely add some noise as the PLL continues to maintain lock on the signal. If there is no way to turn off the CP after the lock bit is set in the tuner, it's probably better to leave it off for lower noise and just live with slower tuning. Leaving the CP bit set should be especially noticable ad FM noise when set to tune to FM radio stations. From the FM1236ME_MK3 datasheet: It is recommended to set CP=0 in the FM mode at all times. But the VHF low band control byte is also used when setting FM radio (AFAICT with a quick look at the code.) Regards, Andy diff -r 43dbc8ebb5a2 linux/drivers/media/common/tuners/tuner-types.c --- a/linux/drivers/media/common/tuners/tuner-types.c Tue Jan 27 23:47:50 2009 -0200 +++ b/linux/drivers/media/common/tuners/tuner-types.c Tue Apr 21 09:44:38 2009 +1000 @@ -557,9 +557,9 @@ /* TUNER_PHILIPS_FM1216ME_MK3 - Philips PAL */ static struct tuner_range tuner_fm1216me_mk3_pal_ranges[] = { - { 16 * 158.00 /*MHz*/, 0x8e, 0x01, }, - { 16 * 442.00 /*MHz*/, 0x8e, 0x02, }, - { 16 * 999.99, 0x8e, 0x04, }, + { 16 * 158.00 /*MHz*/, 0xc6, 0x01, }, + { 16 * 441.00 /*MHz*/, 0xc6, 0x02, }, + { 16 * 864.00, 0xc6, 0x04, }, }; Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov d.beli...@gmail.com With my best regards, Dmitry. Cheers, Hermann -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html